Commit Graph

15 Commits

Author SHA1 Message Date
Sergey Sharybin
89ee6e0808 Cycles: attempt to solve compilation error on 32bit OSX 2014-06-19 12:24:08 +06:00
Sergey Sharybin
3144ae2c34 Cycles: Slight modification to the previous commit
This way util_simd.cpp would not require modifications
if/when SSE2 is suddenly supported on 32bit platforms.

This also allowed to unleash some issues with util_simd.h
related on the fact that there size_t and int are actually
the same types.
2014-06-17 01:00:43 +06:00
Antony Riakiotakis
9073a81bce Fix compilation in cycles 2014-06-14 13:54:38 +03:00
Thomas Dinges
cd5e1ff74e Cycles Refactor: Add SSE Utility code from Embree for cleaner SSE code.
This makes the code a bit easier to understand, and might come in handy
if we want to reuse more Embree code.

Differential Revision: https://developer.blender.org/D482

Code by Brecht, with fixes by Lockal, Sergey and myself.
2014-06-13 21:59:12 +02:00
Campbell Barton
dc13969e48 Style cleanup: indentation, braces 2014-05-05 02:19:08 +10:00
Sv. Lockal
ab32a1807d Cycles: SSE optimization for Voronoi cells texture
Gives 5-6% speedup for Caterpillar_PatazStudio.blend.

Reviewed By: brecht, dingto

Differential Revision: https://developer.blender.org/D419
2014-04-03 23:35:10 +04:00
Sv. Lockal
c45c472e1b Cycles: SSE optimization for line segments/ribbons hair
Gives ~11% speedup for hair.blend, ~10% for koro_final.blend

Also extract few common subexpressions in hair calculation.

Reviewed By: brecht

Differential Revision: https://developer.blender.org/D318
2014-03-23 00:45:59 +04:00
Sv. Lockal
7808360c5f Cycles: fix crash in SSE hair and half-floats on x86+vc2008
MSVC 2008 ignores alignement attribute when assigning from unaligned
float4 vector, returned from other function. Now Cycles uses unaligned
loads instead of casts for win32 in x86 mode.
2014-02-27 15:01:20 +04:00
Sv. Lockal
bd438de8c6 Cycles: bvh_cardinal_curve_intersect intro SSE optimization
Gives 5% speedup for koro_final.blend, 10-15% for hair.blend

Reviewed By: brecht

Differential Revision: https://developer.blender.org/D225
2014-02-04 23:39:02 +04:00
Sv. Lockal
cb96cf0b63 Cycles: small optimization for SSE 4.1 bvh intersector
Gives 0.7% - 1.3% speedup for BMW1M-MikePan scene.

Reviewers: juicyfruit

Differential Revision: https://developer.blender.org/D280
2014-02-03 20:49:07 +04:00
Sv. Lockal
1ff299b317 Cycles: less instructions for CPU perlin code
Also move shuffle() in bvh traversal back (was accidentally lost in SSE hair revert)
2014-02-01 14:07:53 +04:00
Sv. Lockal
47c5898fa1 Cycles: SSE for Voronoi textures (targeted for Haswell CPUs)
Gives up to 15% speedup scenes with voronoi-based textures (up to 25% with volumes) on Haswell. The performance change for other CPUs is much smaller: 1-2%.

Reviewed By: brecht

Differential Revision: https://developer.blender.org/D203
2014-01-12 18:14:00 +04:00
Sv. Lockal
da3fdf0b4b Code Cleanup: in Cycles SSE replace macros with templates, skip unused code with preprocessor, simplify casts 2014-01-11 22:20:03 +04:00
Sv. Lockal
96903508bc Cycles: SSE optimization for sRGB conversion (gives 7% speedup on CPU for pavillon_barcelone scene)
Thanks brecht/dingto/juicyfruit et al. for testing and reviewing this patch in T38034.
2014-01-06 20:03:30 +04:00
Thomas Dinges
1578b55c27 Cycles: Move SIMD utility functions into its own file.
Recently added SSE macros for noise texture can be moved here as well, but I leave this for later.
2013-12-27 21:30:21 +01:00